Product Owner (for AI Delivery)

Job Family

Product Owner AIEnabled Products

The Product Owner represents the business within the squad and acts as the primary interface for demand intake including AIenabled datadriven and automation use cases.

The Product Owner reflects accepted demand on the squad backlog and prioritizes it according to business value strategic priorities regulatory constraints (banking license) and AI risk posture.

Role Purpose

The Product Owner is the guardian of product fitness for purpose ensuring that functional nonfunctional and AIspecific requirements are met for products of limited complexity uncertainty and dependencies (e.g. mature products endoflife systems or products with a welldefined operational scope).

This includes ensuring that AI components (models data pipelines decision logic automation) are:

• Fit for business intent
• Compliant with regulatory and ethical standards
• Operationally robust and explainable

Description & Responsibilities

1. Product Ownership & Business Value (AIAware)

• Act asendtoend owner of the product including:
  • Functional requirements
  • Nonfunctional requirements (performance security resilience)
  • AIspecific qualities such as explainability data quality bias awareness and model lifecycle sustainability
• Linkbusiness value to the Product Backlog explicitly identifying:
  • Where AI or automation contributes to efficiency risk reduction or customer value
  • WherenonAI solutions are preferable ensuring pragmatic and valuedriven decisions
• Represent thebusiness intent behind AI usage ensuring the squad understands:
  • Why AI is used
  • What decisions it supports or automates
  • What human oversight is required

2. Stakeholder & Customer Centricity (AI Context)

• Identify and manage stakeholders (business sponsors operations risk (EU AI Risks associated as well) compliance legal IT data architecture).
• Collect and federate stakeholder input on:
  • Business outcomes
  • Regulatory constraints
  • AI acceptability (risk appetite explainability auditability)
• Guide the squad towardscustomercentric and usercentric AI solutions ensuring:
  • Transparency of AIdriven decisions
  • Clear communication of AI limitations and confidence levels

3. Backlog Management & Story Definition (AIReady)

• Own and manage theProduct Backlog ensuring it is:
  • Complete transparent prioritized and understood
  • Inclusive ofAI lifecycle work not just features
• Effectively write and slice stories that may include:
  • Data sourcing and preparation
  • Feature engineering (transforming raw data into modelready inputs)
  • Model inference integration (how predictions are consumed by systems)
  • Humanintheloop controls (human validation or override of AI outputs)
• Ensure stories includeAIrelevant acceptance criteria such as:
  • Accuracy or quality thresholds
  • Explainability requirements
  • Monitoring and logging expectations

4. Collaboration with Epic Owner & TPO (AI Alignment)

(TPO Technical Product Owner responsible for technical coherence)

• Work closely with theEpic Owner andTPO to:
  • Maximize business value from AI and data capabilities
  • Align AI initiatives with strategic priorities at epic and feature level
  • Coown business objectives including AIenabled outcomes
• Refine Features into Product Backlog Items (PBIs) that reflect:
  • Business intent
  • Technical feasibility
  • AI risk and compliance constraints

5. Delivery Oversight & Risk Management (AI & Regulatory)

• Oversee delivery stages and ensureall risks are identified and mitigated including:
  • Regulatory risks (e.g. CSDR Central Securities Depositories Regulation)
  • Compliance and data protection (e.g. GDPR General Data Protection Regulation)
  • Security and architecture risks
  • AIspecific risks:
    • Model bias
    • Lack of explainability
    • Data drift (changes in data patterns over time)
    • Model drift (degradation of model performance in production)
• Ensure AI solutions comply with:
  • Internal AI governance frameworks
  • Model risk management expectations
  • Audit and traceability requirements

6. Sprint Execution & Value Validation

• Define with the squad:
  • Sprint goals
  • Sprint content
  • Readiness of AIrelated work (data availability environments dependencies)
• Facilitate sprint reviews and demonstrations ensuring:
  • AI outcomes are explained in business terms
  • Limitations and confidence levels are transparently communicated
• Validate and accept or reject delivered stories and features including:
  • Verification that AI outputs meet agreed acceptance criteria
  • Confirmation that monitoring and controls are in place

7. Measurement KPIs & Continuous Improvement (AIInformed)

• Define and pilotProduct and Business KPIs with support from senior colleagues including:
  • Traditional KPIs (throughput adoption value delivered)
  • AIspecific indicators such as:
    • Prediction quality trends
    • Automation rates vs. manual intervention
    • Exception and override frequency
• Actively collect feedback from the squad and stakeholders and translate it into backlog improvements.
• Assess and demonstrate value delivered at squad level (e.g. squad health check boards) ensuring AI contributions aremeasurable and defensible.

Role Scope & Support

• Operates on products oflimited complexity uncertainty and dependencies such as:
  • Mature or endoflife products
  • Welldefined operational scopes
  • AI components with controlled impact and clear governance
• Receives guidance from senior colleagues for:
  • Strategic decisions
  • Complex prioritization tradeoffs
  • AIrelated risk or compliance decisions

Key Competencies (AIInfused)

• Strong Product Ownership fundamentals (Agile backlog management value prioritization)
• AI and data literacy including:
  • Understanding of the AI lifecycle (data model deployment monitoring)
  • Ability to translate business needs into AIready requirements
• Awareness ofAI governance compliance and ethical considerations
• Ability to collaborate effectively with:
  • Data Scientists
  • Machine Learning Engineers
  • Architects and Risk/Compliance stakeholders

Final Note (Positioning)

This role does not require handson model building but it does require sufficient AI technology stack understanding to:

• Ask the right questions
• Prioritize the right work
• Ensure AI deliversreal compliant and sustainable business value